selftests/mm: assert rmap behave as expected
Subsystems (4)
cover selftests/mm: assert rmap behave as expected
As David suggested, currently we don't have a high level test case to verify the behavior of rmap. This patch set introduce the verification on rmap by migration.
Patch 1 is a trivial cleanup.
Patch 2 is a preparation to move ksm related operation into vm_util. Patch 3 is the new test case.
Currently it covers following four scenarios:
* anonymous page
* shmem page
* pagecache page
* ksm page
RFC->v1:
* open file in function itself instead of pass fd as paremeter
* fault in the region by accessing it instead of print content
Wei Yang (3):
selftests/mm: check a valid fd with negative value selftests/mm: put general ksm operation into vm_util selftests/mm: assert rmap behave as expected
MAINTAINERS | 1 + tools/testing/selftests/mm/.gitignore | 1 + tools/testing/selftests/mm/Makefile | 3 + .../selftests/mm/ksm_functional_tests.c | 95 +--- tools/testing/selftests/mm/rmap.c | 470 ++++++++++++++++++ tools/testing/selftests/mm/run_vmtests.sh | 4 + tools/testing/selftests/mm/vm_util.c | 111 +++++ tools/testing/selftests/mm/vm_util.h | 7 + 8 files changed, 605 insertions(+), 87 deletions(-) create mode 100644 tools/testing/selftests/mm/rmap.c
Signature · 1 line
2.34.1
PATCH 1/3 selftests/mm: check a valid fd with negative value
fd 0 is usually occupied by stdin, which is not expected to be returned by open(). But 0 is still a valid fd.
Check the valid fd with negative value like other places.
--- tools/testing/selftests/mm/ksm_functional_tests.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/mm/ksm_functional_tests.c +1 -1
@@ -81,7 +81,7 @@ static long get_my_ksm_zero_pages(void)ssize_t read_size;unsigned long my_ksm_zero_pages; -if (!proc_self_ksm_stat_fd) +if (proc_self_ksm_stat_fd <0 )return 0 ; read_size =pread (proc_self_ksm_stat_fd, buf,sizeof (buf) -1 ,0 );
Signature · 1 line
2.34.1
Re: [PATCH 1/3] selftests/mm: check a valid fd with negative value
Quoted 21 lines · Wei Yang
fd 0 is usually occupied by stdin, which is not expected to be returned by
open(). But 0 is still a valid fd.Check the valid fd with negative value like other places.
Signed-off-by Wei Yang <[email protected]>Cc David Hildenbrand <[email protected]>Cc Lorenzo Stoakes <[email protected]>Cc Rik van Riel <[email protected]>Cc Liam R. Howlett <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Harry Yoo <[email protected]>--- tools/testing/selftests/mm/ksm_functional_tests.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)tools/testing/selftests/mm/ksm_functional_tests.c +0
-0
tools/testing/selftests/mm/ksm_functional_tests.c static long get_my_ksm_zero_pages(void) · line 85
Quoted 6 lines
tools/testing/selftests/mm/ksm_functional_tests.c +1
-1@@ -81,7 +81,7 @@ static long get_my_ksm_zero_pages(void)ssize_t read_size;unsigned long my_ksm_zero_pages; -if (!proc_self_ksm_stat_fd) +if (proc_self_ksm_stat_fd <0 )return 0 ;
Only test_unmerge_zero_pages() calls get_my_ksm_zero_pages(). And there, we have
if (proc_self_ksm_stat_fd < 0) {
ksft_test_result_skip("open(\"/proc/self/ksm_stat\") failed\n");
return;
}As the fd is opened after others, we should never get fd=0.
So probably, this should just be handled as part of factoring it out, no need for this change upfront AFAIKS.
Signature · 3 lines
Cheers, David / dhildenb
Re: [PATCH 1/3] selftests/mm: check a valid fd with negative value
Quoted 31 lines · David Hildenbrand
Quoted 21 lines · Wei Yang
fd 0 is usually occupied by stdin, which is not expected to be returned by
open(). But 0 is still a valid fd.Check the valid fd with negative value like other places.
Signed-off-by Wei Yang <[email protected]>Cc David Hildenbrand <[email protected]>Cc Lorenzo Stoakes <[email protected]>Cc Rik van Riel <[email protected]>Cc Liam R. Howlett <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Harry Yoo <[email protected]>--- tools/testing/selftests/mm/ksm_functional_tests.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)tools/testing/selftests/mm/ksm_functional_tests.c +0
-0tools/testing/selftests/mm/ksm_functional_tests.c static long get_my_ksm_zero_pages(void) · line 84
Quoted 5 lines
tools/testing/selftests/mm/ksm_functional_tests.c +1
-1@@ -81,7 +81,7 @@ static long get_my_ksm_zero_pages(void)ssize_t read_size;unsigned long my_ksm_zero_pages; -if (!proc_self_ksm_stat_fd) +if (proc_self_ksm_stat_fd <0 )return 0 ;Only
test_unmerge_zero_pages()callsget_my_ksm_zero_pages(). And there, we have
Quoted 7 lines
if (proc_self_ksm_stat_fd < 0) { ksft_test_result_skip("open(\"/proc/self/ksm_stat\") failed\n"); return; }As the fd is opened after others, we should never get fd=0.
You are right.
Quoted 3 lines
So probably, this should just be handled as part of factoring it out, no need for this change upfront AFAIKS.
My intention of this change is, we plan to move get_my_ksm_zero_pages() into a standalone helper and will open and check the fd there. To make following change consistent, I split it out.
If you think it is not necessary, will drop this.
Quoted 4 lines
Signature · 3 lines
Cheers, David / dhildenb
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH 1/3] selftests/mm: check a valid fd with negative value
Quoted 46 lines · Wei Yang
Quoted 31 lines · David Hildenbrand
Quoted 21 lines · Wei Yang
fd 0 is usually occupied by stdin, which is not expected to be returned by
open(). But 0 is still a valid fd.Check the valid fd with negative value like other places.
Signed-off-by Wei Yang <[email protected]>Cc David Hildenbrand <[email protected]>Cc Lorenzo Stoakes <[email protected]>Cc Rik van Riel <[email protected]>Cc Liam R. Howlett <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Harry Yoo <[email protected]>--- tools/testing/selftests/mm/ksm_functional_tests.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)tools/testing/selftests/mm/ksm_functional_tests.c +0
-0tools/testing/selftests/mm/ksm_functional_tests.c static long get_my_ksm_zero_pages(void) · line 84
Quoted 5 lines
tools/testing/selftests/mm/ksm_functional_tests.c +1
-1@@ -81,7 +81,7 @@ static long get_my_ksm_zero_pages(void)ssize_t read_size;unsigned long my_ksm_zero_pages; -if (!proc_self_ksm_stat_fd) +if (proc_self_ksm_stat_fd <0 )return 0 ;Only
test_unmerge_zero_pages()callsget_my_ksm_zero_pages(). And there, we haveQuoted 7 lines
if (proc_self_ksm_stat_fd < 0) { ksft_test_result_skip("open(\"/proc/self/ksm_stat\") failed\n"); return; }As the fd is opened after others, we should never get fd=0.
You are right.
Quoted 3 lines
So probably, this should just be handled as part of factoring it out, no need for this change upfront AFAIKS.
Quoted 5 lines
My intention of this change is, we plan to move
get_my_ksm_zero_pages()into a standalone helper and will open and check the fd there. To make following change consistent, I split it out.If you think it is not necessary, will drop this.
IMHO, just cleaning it up while factoring out is better.
Signature · 3 lines
Cheers, David / dhildenb
PATCH 2/3 selftests/mm: put general ksm operation into vm_util
In reply to Wei YangThere are some general ksm operations could be used by other related test case. Put them into vm_util for common use.
This is a preparation patch for later use.
---
v1: open/close fd in function itself instead of pass as parameter
--- .../selftests/mm/ksm_functional_tests.c | 95 ++------------- tools/testing/selftests/mm/vm_util.c | 111 ++++++++++++++++++ tools/testing/selftests/mm/vm_util.h | 7 ++ 3 files changed, 126 insertions(+), 87 deletions(-)
tools/testing/selftests/mm/ksm_functional_tests.c +8 -87
@@ -38,11 +38,6 @@ enum ksm_merge_mode { };static int mem_fd; -static int ksm_fd; -static int ksm_full_scans_fd; -static int proc_self_ksm_stat_fd; -static int proc_self_ksm_merging_pages_fd; -static int ksm_use_zero_pages_fd;static int pagemap_fd;static size_t pagesize;
@@ -73,62 +68,6 @@ static bool range_maps_duplicates(char *addr, unsigned long size)return false; } -static long get_my_ksm_zero_pages (void ) -{ -char buf[200 ]; -char *substr_ksm_zero; -size_t value_pos; -ssize_t read_size; -unsigned long my_ksm_zero_pages; - -if (proc_self_ksm_stat_fd <0 ) -return 0 ; - - read_size =pread (proc_self_ksm_stat_fd, buf,sizeof (buf) -1 ,0 ); -if (read_size <0 ) -return -errno; - - buf[read_size] =0 ; - - substr_ksm_zero =strstr (buf,"ksm_zero_pages" ); -if (!substr_ksm_zero) -return 0 ; - - value_pos =strcspn (substr_ksm_zero,"0123456789" ); - my_ksm_zero_pages =strtol (substr_ksm_zero + value_pos,NULL ,10 ); - -return my_ksm_zero_pages; -} - -static long get_my_merging_pages (void ) -{ -char buf[10 ]; -ssize_t ret; - -if (proc_self_ksm_merging_pages_fd <0 ) -return proc_self_ksm_merging_pages_fd; - - ret =pread (proc_self_ksm_merging_pages_fd, buf,sizeof (buf) -1 ,0 ); -if (ret <=0 ) -return -errno; - buf[ret] =0 ; - -return strtol (buf,NULL ,10 ); -} - -static long ksm_get_full_scans (void ) -{ -char buf[10 ]; -ssize_t ret; - - ret =pread (ksm_full_scans_fd, buf,sizeof (buf) -1 ,0 ); -if (ret <=0 ) -return -errno; - buf[ret] =0 ; - -return strtol (buf,NULL ,10 ); -} -static int ksm_merge (void ) {long start_scans, end_scans;
@@ -137,7 +76,7 @@ static int ksm_merge(void) start_scans =ksm_get_full_scans ();if (start_scans <0 )return start_scans; -if (write (ksm_fd,"1" ,1 ) !=1 ) +if (ksm_start_and_merge ())return -errno; do { end_scans = ksm_get_full_scans();
@@ -150,7 +89,7 @@ static int ksm_merge(void)static int ksm_unmerge (void ) { -if (write (ksm_fd,"2" ,1 ) !=1 ) +if (ksm_stop_and_unmerge ())return -errno;return 0 ; }
@@ -168,7 +107,7 @@ static char *__mmap_and_merge_range(char val, unsigned long size, int prot,return err_map; } -if (get_my_merging_pages () >0 ) { +if (ksm_get_self_merging_pages () >0 ) {ksft_print_msg ("Still pages merged\n" );return err_map; }
@@ -227,7 +166,7 @@ static char *__mmap_and_merge_range(char val, unsigned long size, int prot,* Check if anything was merged at all. Ignore the zero page that is * accounted differently (depending on kernel support). */ -if (val && !get_my_merging_pages ()) { +if (val && !ksm_get_self_merging_pages ()) {ksft_print_msg ("No pages got merged\n" ); goto unmap; }
@@ -286,15 +225,7 @@ static void test_unmerge_zero_pages(void)ksft_print_msg ("[RUN] %s\n" , __func__); -if (proc_self_ksm_stat_fd <0 ) { -ksft_test_result_skip ("open(\"/proc/self/ksm_stat\") failed\n" ); -return ; - } -if (ksm_use_zero_pages_fd <0 ) { -ksft_test_result_skip ("open \"/sys/kernel/mm/ksm/use_zero_pages\" failed\n" ); -return ; - } -if (write (ksm_use_zero_pages_fd,"1" ,1 ) !=1 ) { +if (ksm_use_zero_pages ()) {ksft_test_result_skip ("write \"/sys/kernel/mm/ksm/use_zero_pages\" failed\n" );return ; }
@@ -306,7 +237,7 @@ static void test_unmerge_zero_pages(void)/* Check if ksm_zero_pages is updated correctly after KSM merging */ pages_expected = size / pagesize; -if (pages_expected !=get_my_ksm_zero_pages ()) { +if (pages_expected !=ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after merging\n" ); goto unmap; }
@@ -319,7 +250,7 @@ static void test_unmerge_zero_pages(void)/* Check if ksm_zero_pages is updated correctly after unmerging */ pages_expected /=2 ; -if (pages_expected !=get_my_ksm_zero_pages ()) { +if (pages_expected !=ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after unmerging\n" ); goto unmap; }
@@ -329,7 +260,7 @@ static void test_unmerge_zero_pages(void) *((unsigned int *)&map[offs]) = offs;/* Now we should have no zeropages remaining. */ -if (get_my_ksm_zero_pages ()) { +if (ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after write fault\n" ); goto unmap; }
@@ -685,19 +616,9 @@ static void init_global_file_handles(void) mem_fd =open ("/proc/self/mem" ,O_RDWR );if (mem_fd <0 )ksft_exit_fail_msg ("opening /proc/self/mem failed\n" ); - ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR ); -if (ksm_fd <0 ) -ksft_exit_skip ("open(\"/sys/kernel/mm/ksm/run\") failed\n" ); - ksm_full_scans_fd =open ("/sys/kernel/mm/ksm/full_scans" ,O_RDONLY ); -if (ksm_full_scans_fd <0 ) -ksft_exit_skip ("open(\"/sys/kernel/mm/ksm/full_scans\") failed\n" ); pagemap_fd =open ("/proc/self/pagemap" ,O_RDONLY );if (pagemap_fd <0 )ksft_exit_skip ("open(\"/proc/self/pagemap\") failed\n" ); - proc_self_ksm_stat_fd =open ("/proc/self/ksm_stat" ,O_RDONLY ); - proc_self_ksm_merging_pages_fd =open ("/proc/self/ksm_merging_pages" , -O_RDONLY ); - ksm_use_zero_pages_fd =open ("/sys/kernel/mm/ksm/use_zero_pages" ,O_RDWR ); }int main (int argc,char **argv)
tools/testing/selftests/mm/vm_util.c +111 -0
114 lines not on this page · raw message
tools/testing/selftests/mm/vm_util.h +7 -0
@@ -122,6 +122,13 @@ static inline void log_test_result(int result)void *sys_mremap (void *old_address,unsigned long old_size,unsigned long new_size,int flags,void *new_address); +long ksm_get_self_zero_pages (); +long ksm_get_self_merging_pages (); +long ksm_get_full_scans (); +int ksm_use_zero_pages (); +int ksm_start_and_merge (); +int ksm_stop_and_unmerge (); +/* * On ppc64 this will only work with radix 2M hugepage size */
Signature · 1 line
2.34.1
Re: [PATCH 2/3] selftests/mm: put general ksm operation into vm_util
Quoted 89 lines · Wei Yang
There are some general ksm operations could be used by other related test case. Put them into
vm_utilfor common use.This is a preparation patch for later use.
Signed-off-by Wei Yang <[email protected]>Suggested-by David Hildenbrand <[email protected]>Cc David Hildenbrand <[email protected]>Cc Lorenzo Stoakes <[email protected]>Cc Rik van Riel <[email protected]>Cc Liam R. Howlett <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Harry Yoo <[email protected]>---
v1: open/close fd in function itself instead of pass as parameter--- .../selftests/mm/ksm_functional_tests.c | 95 ++------------- tools/testing/selftests/mm/vm_util.c | 111 ++++++++++++++++++ tools/testing/selftests/mm/vm_util.h | 7 ++ 3 files changed, 126 insertions(+), 87 deletions(-)tools/testing/selftests/mm/ksm_functional_tests.c +0
-52@@ -38,11 +38,6 @@ enum ksm_merge_mode { };static int mem_fd; -static int ksm_fd; -static int ksm_full_scans_fd; -static int proc_self_ksm_stat_fd; -static int proc_self_ksm_merging_pages_fd; -static int ksm_use_zero_pages_fd;static int pagemap_fd;static size_t pagesize;@@ -73,62 +68,6 @@ static bool range_maps_duplicates(char *addr, unsigned long size)return false; } -static long get_my_ksm_zero_pages (void ) -{ -char buf[200 ]; -char *substr_ksm_zero; -size_t value_pos; -ssize_t read_size; -unsigned long my_ksm_zero_pages; - -if (proc_self_ksm_stat_fd <0 ) -return 0 ; - - read_size =pread (proc_self_ksm_stat_fd, buf,sizeof (buf) -1 ,0 ); -if (read_size <0 ) -return -errno; - - buf[read_size] =0 ; - - substr_ksm_zero =strstr (buf,"ksm_zero_pages" ); -if (!substr_ksm_zero) -return 0 ; - - value_pos =strcspn (substr_ksm_zero,"0123456789" ); - my_ksm_zero_pages =strtol (substr_ksm_zero + value_pos,NULL ,10 ); - -return my_ksm_zero_pages; -} - -static long get_my_merging_pages (void ) -{ -char buf[10 ]; -ssize_t ret; - -if (proc_self_ksm_merging_pages_fd <0 ) -return proc_self_ksm_merging_pages_fd; - - ret =pread (proc_self_ksm_merging_pages_fd, buf,sizeof (buf) -1 ,0 ); -if (ret <=0 ) -return -errno; - buf[ret] =0 ; - -return strtol (buf,NULL ,10 ); -} - -static long ksm_get_full_scans (void ) -{ -char buf[10 ]; -ssize_t ret;
tools/testing/selftests/mm/ksm_functional_tests.c static bool range_maps_duplicates(char *addr, unsigned long size) · line 71
Quoted 10 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-9- - ret =pread (ksm_full_scans_fd, buf,sizeof (buf) -1 ,0 ); -if (ret <=0 ) -return -errno; - buf[ret] =0 ; - -return strtol (buf,NULL ,10 ); -} -static int ksm_merge (void )
Should ksm_merge() get factored out as well?
Quoted 12 lines
tools/testing/selftests/mm/ksm_functional_tests.c +1
-1{long start_scans, end_scans;@@ -137,7 +76,7 @@ static int ksm_merge(void) start_scans =ksm_get_full_scans ();if (start_scans <0 )return start_scans; -if (write (ksm_fd,"1" ,1 ) !=1 ) +if (ksm_start_and_merge ())return -errno; do { end_scans = ksm_get_full_scans();
tools/testing/selftests/mm/ksm_functional_tests.c static int ksm_merge(void) · line 94
Quoted 7 lines
tools/testing/selftests/mm/ksm_functional_tests.c +1
-1@@ -150,7 +89,7 @@ static int ksm_merge(void)static int ksm_unmerge (void ) { -if (write (ksm_fd,"2" ,1 ) !=1 ) +if (ksm_stop_and_unmerge ())return -errno;return 0 ;
What's the reason of gaving ksm_unmerge() and ksm_stop_and_unmerge()?
Probably we should just use ksm_stop_and_unmerge() and remove ksm_unmerge().
See below regarding letting ksm_stop_and_unmerge() and friends return -errno in case of error.
Quoted 20 lines
tools/testing/selftests/mm/ksm_functional_tests.c +2
-2}@@ -168,7 +107,7 @@ static char *__mmap_and_merge_range(char val, unsigned long size, int prot,return err_map; } -if (get_my_merging_pages () >0 ) { +if (ksm_get_self_merging_pages () >0 ) {ksft_print_msg ("Still pages merged\n" );return err_map; }@@ -227,7 +166,7 @@ static char *__mmap_and_merge_range(char val, unsigned long size, int prot,* Check if anything was merged at all. Ignore the zero page that is * accounted differently (depending on kernel support). */ -if (val && !get_my_merging_pages ()) { +if (val && !ksm_get_self_merging_pages ()) {ksft_print_msg ("No pages got merged\n" ); goto unmap; }
tools/testing/selftests/mm/ksm_functional_tests.c static void test_unmerge_zero_pages(void) · line 292
Quoted 7 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-4@@ -286,15 +225,7 @@ static void test_unmerge_zero_pages(void)ksft_print_msg ("[RUN] %s\n" , __func__); -if (proc_self_ksm_stat_fd <0 ) { -ksft_test_result_skip ("open(\"/proc/self/ksm_stat\") failed\n" ); -return ; - }
See below: probably we should do a test read so we know the file exists and can be read. So we don't get misleading errors later on older kernels.
if (ksm_get_self_zero_pages() < 0) {
ksft_test_result_skip("accessing \"/proc/self/ksm_stat\" failed\n");
return;
}Quoted 37 lines
tools/testing/selftests/mm/ksm_functional_tests.c +4
-8-if (ksm_use_zero_pages_fd <0 ) { -ksft_test_result_skip ("open \"/sys/kernel/mm/ksm/use_zero_pages\" failed\n" ); -return ; - } -if (write (ksm_use_zero_pages_fd,"1" ,1 ) !=1 ) { +if (ksm_use_zero_pages ()) {ksft_test_result_skip ("write \"/sys/kernel/mm/ksm/use_zero_pages\" failed\n" );return ; }@@ -306,7 +237,7 @@ static void test_unmerge_zero_pages(void)/* Check if ksm_zero_pages is updated correctly after KSM merging */ pages_expected = size / pagesize; -if (pages_expected !=get_my_ksm_zero_pages ()) { +if (pages_expected !=ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after merging\n" ); goto unmap; }@@ -319,7 +250,7 @@ static void test_unmerge_zero_pages(void)/* Check if ksm_zero_pages is updated correctly after unmerging */ pages_expected /=2 ; -if (pages_expected !=get_my_ksm_zero_pages ()) { +if (pages_expected !=ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after unmerging\n" ); goto unmap; }@@ -329,7 +260,7 @@ static void test_unmerge_zero_pages(void) *((unsigned int *)&map[offs]) = offs;/* Now we should have no zeropages remaining. */ -if (get_my_ksm_zero_pages ()) { +if (ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after write fault\n" ); goto unmap; }
tools/testing/selftests/mm/ksm_functional_tests.c static void init_global_file_handles(void) · line 693
Quoted 9 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-6@@ -685,19 +616,9 @@ static void init_global_file_handles(void) mem_fd =open ("/proc/self/mem" ,O_RDWR );if (mem_fd <0 )ksft_exit_fail_msg ("opening /proc/self/mem failed\n" ); - ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR ); -if (ksm_fd <0 ) -ksft_exit_skip ("open(\"/sys/kernel/mm/ksm/run\") failed\n" ); - ksm_full_scans_fd =open ("/sys/kernel/mm/ksm/full_scans" ,O_RDONLY ); -if (ksm_full_scans_fd <0 ) -ksft_exit_skip ("open(\"/sys/kernel/mm/ksm/full_scans\") failed\n" );
For these skip cases, we should probably do a test access.
if (ksm_stop_and_unmerge() < 0)
ksft_exit_skip("accessing \"/sys/kernel/mm/ksm/run\\") failed\n");
if (ksm_get_full_scans() < 0)
ksft_exit_skip("accessing \"/sys/kernel/mm/ksm/full_scans\") failed\n");So we later now that the files actually do exist and can be opened+written.
Quoted 14 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-4pagemap_fd = open("/proc/self/pagemap", O_RDONLY);> if (pagemap_fd < 0)ksft_exit_skip ("open(\"/proc/self/pagemap\") failed\n" ); - proc_self_ksm_stat_fd =open ("/proc/self/ksm_stat" ,O_RDONLY ); - proc_self_ksm_merging_pages_fd =open ("/proc/self/ksm_merging_pages" , -O_RDONLY ); - ksm_use_zero_pages_fd =open ("/sys/kernel/mm/ksm/use_zero_pages" ,O_RDWR ); }int main (int argc,char **argv)tools/testing/selftests/mm/vm_util.c +0
-0
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 559
Quoted 5 lines
tools/testing/selftests/mm/vm_util.c +2
-0@@ -555,3 +555,114 @@ bool detect_huge_zeropage(void)close (fd);return enabled; } + +long ksm_get_self_zero_pages ()
For all these functions without parameters:
(void)
Quoted 66 lines
tools/testing/selftests/mm/vm_util.c +66
-0+{ +int proc_self_ksm_stat_fd; +char buf[200 ]; +char *substr_ksm_zero; +size_t value_pos; +ssize_t read_size; +unsigned long my_ksm_zero_pages; + + proc_self_ksm_stat_fd =open ("/proc/self/ksm_stat" ,O_RDONLY ); +if (proc_self_ksm_stat_fd <0 ) +return 0 ; + + read_size =pread (proc_self_ksm_stat_fd, buf,sizeof (buf) -1 ,0 ); +close (proc_self_ksm_stat_fd); +if (read_size <0 ) +return -errno; + + buf[read_size] =0 ; + + substr_ksm_zero =strstr (buf,"ksm_zero_pages" ); +if (!substr_ksm_zero) +return 0 ; + + value_pos =strcspn (substr_ksm_zero,"0123456789" ); + my_ksm_zero_pages =strtol (substr_ksm_zero + value_pos,NULL ,10 ); + +return my_ksm_zero_pages; +} + +long ksm_get_self_merging_pages () +{ +int proc_self_ksm_merging_pages_fd; +char buf[10 ]; +ssize_t ret; + + proc_self_ksm_merging_pages_fd =open ("/proc/self/ksm_merging_pages" , +O_RDONLY ); +if (proc_self_ksm_merging_pages_fd <0 ) +return proc_self_ksm_merging_pages_fd; + + ret =pread (proc_self_ksm_merging_pages_fd, buf,sizeof (buf) -1 ,0 ); +close (proc_self_ksm_merging_pages_fd); +if (ret <=0 ) +return -errno; + buf[ret] =0 ; + +return strtol (buf,NULL ,10 ); +} + +long ksm_get_full_scans () +{ +int ksm_full_scans_fd; +char buf[10 ]; +ssize_t ret; + + ksm_full_scans_fd =open ("/sys/kernel/mm/ksm/full_scans" ,O_RDONLY ); +if (ksm_full_scans_fd <0 ) +return ksm_full_scans_fd; + + ret =pread (ksm_full_scans_fd, buf,sizeof (buf) -1 ,0 ); +close (ksm_full_scans_fd); +if (ret <=0 ) +return -errno; + buf[ret] =0 ; + +return strtol (buf,NULL ,10 );
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 635
Quoted 10 lines
tools/testing/selftests/mm/vm_util.c +10
-0+} + +int ksm_use_zero_pages () +{ +int ksm_use_zero_pages_fd; +ssize_t ret; + + ksm_use_zero_pages_fd =open ("/sys/kernel/mm/ksm/use_zero_pages" ,O_RDWR ); +if (ksm_use_zero_pages_fd <0 ) +return -1 ;
For all these functions, we should not require the caller to lookup errno but instead return it.
return -errno;
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 639
Quoted 4 lines
tools/testing/selftests/mm/vm_util.c +4
-0+ + ret =write (ksm_use_zero_pages_fd,"1" ,1 ); +close (ksm_use_zero_pages_fd); +return ret ==1 ?0 : ret;
return ret == 1 ? 0 : -errno;
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 649
Quoted 10 lines
tools/testing/selftests/mm/vm_util.c +10
-0+} + +int ksm_start_and_merge () +{ +int ksm_fd; +ssize_t ret; + + ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR ); +if (ksm_fd <0 ) +return -1 ;
Same comments regarding errno.
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 653
Quoted 4 lines
tools/testing/selftests/mm/vm_util.c +4
-0+ + ret =write (ksm_fd,"1" ,1 ); +close (ksm_fd); +return ret ==1 ?0 : ret;
double space
Quoted 4 lines
tools/testing/selftests/mm/vm_util.c +4
-0+} + +int ksm_stop_and_unmerge () +{
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 667
Quoted 10 lines
tools/testing/selftests/mm/vm_util.c +10
-0+int ksm_fd; +ssize_t ret; + + ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR ); +if (ksm_fd <0 ) +return -1 ; + + ret =write (ksm_fd,"2" ,1 ); +close (ksm_fd); +return ret ==1 ?0 : ret;
double space
Signature · 3 lines
Cheers, David / dhildenb
Re: [PATCH 2/3] selftests/mm: put general ksm operation into vm_util
Thanks for the detailed comment.
Quoted 1 line · David Hildenbrand
On 16.07.25 10:27, Wei Yang wrote:
[...]
Quoted 8 lines
Quoted 8 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-8-static long ksm_get_full_scans (void ) -{ -char buf[10 ]; -ssize_t ret; - - ret =pread (ksm_full_scans_fd, buf,sizeof (buf) -1 ,0 ); -if (ret <=0 ) -return -errno;
Quoted 9 lines
tools/testing/selftests/mm/ksm_functional_tests.c static bool range_maps_duplicates(char *addr, unsigned long size) · line 71
Quoted 6 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-5- buf[ret] =0 ; - -return strtol (buf,NULL ,10 ); -} -static int ksm_merge (void )Should
ksm_merge()get factored out as well?
Reasonable, will factor out.
While one thing interesting is if it scan 2 full round, the ksm case in patch 3 would fail sometimes. But I don't see the failure in ksm_functional_test here.
Quoted 12 lines
Quoted 12 lines
{ long start_scans, end_scans;tools/testing/selftests/mm/ksm_functional_tests.c +1
-1@@ -137,7 +76,7 @@ static int ksm_merge(void) start_scans =ksm_get_full_scans ();if (start_scans <0 )return start_scans; -if (write (ksm_fd,"1" ,1 ) !=1 ) +if (ksm_start_and_merge ())return -errno; do { end_scans = ksm_get_full_scans();@@ -150,7 +89,7 @@ static int ksm_merge(void)
Quoted 9 lines
tools/testing/selftests/mm/ksm_functional_tests.c static int ksm_merge(void) · line 94
Quoted 6 lines
tools/testing/selftests/mm/ksm_functional_tests.c +1
-1static int ksm_unmerge (void ) { -if (write (ksm_fd,"2" ,1 ) !=1 ) +if (ksm_stop_and_unmerge ())return -errno;return 0 ;What's the reason of gaving
ksm_unmerge()andksm_stop_and_unmerge()?
My plan is there are two pairs of helper:
ksm_merge() <-> ksm_unmerge()
ksm_start_and_merge() <-> ksm_stop_and_unmerge()Quoted 2 lines
Probably we should just use
ksm_stop_and_unmerge()and removeksm_unmerge().
Looks reasonable, will remove it.
So how about leave three helpers:
ksm_merge()
ksm_start()
ksm_stop_and_unmerge()
Would this be better?
Quoted 30 lines
See below regarding letting
ksm_stop_and_unmerge()and friends return -errno in case of error.Quoted 19 lines
tools/testing/selftests/mm/ksm_functional_tests.c +2
-2}@@ -168,7 +107,7 @@ static char *__mmap_and_merge_range(char val, unsigned long size, int prot,return err_map; } -if (get_my_merging_pages () >0 ) { +if (ksm_get_self_merging_pages () >0 ) {ksft_print_msg ("Still pages merged\n" );return err_map; }@@ -227,7 +166,7 @@ static char *__mmap_and_merge_range(char val, unsigned long size, int prot,* Check if anything was merged at all. Ignore the zero page that is * accounted differently (depending on kernel support). */ -if (val && !get_my_merging_pages ()) { +if (val && !ksm_get_self_merging_pages ()) {ksft_print_msg ("No pages got merged\n" ); goto unmap; }tools/testing/selftests/mm/ksm_functional_tests.c static void test_unmerge_zero_pages(void) · line 290
Quoted 5 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-4@@ -286,15 +225,7 @@ static void test_unmerge_zero_pages(void)ksft_print_msg ("[RUN] %s\n" , __func__); -if (proc_self_ksm_stat_fd <0 ) { -ksft_test_result_skip ("open(\"/proc/self/ksm_stat\") failed\n" ); -return ; - }See below: probably we should do a test read so we know the file exists and can be read. So we don't get misleading errors later on older kernels.
Quoted 4 lines
if (ksm_get_self_zero_pages() < 0) {
ksft_test_result_skip("accessing \"/proc/self/ksm_stat\" failed\n");
return;
}Agree
Quoted 47 lines
Quoted 35 lines
tools/testing/selftests/mm/ksm_functional_tests.c +4
-8-if (ksm_use_zero_pages_fd <0 ) { -ksft_test_result_skip ("open \"/sys/kernel/mm/ksm/use_zero_pages\" failed\n" ); -return ; - } -if (write (ksm_use_zero_pages_fd,"1" ,1 ) !=1 ) { +if (ksm_use_zero_pages ()) {ksft_test_result_skip ("write \"/sys/kernel/mm/ksm/use_zero_pages\" failed\n" );return ; }@@ -306,7 +237,7 @@ static void test_unmerge_zero_pages(void)/* Check if ksm_zero_pages is updated correctly after KSM merging */ pages_expected = size / pagesize; -if (pages_expected !=get_my_ksm_zero_pages ()) { +if (pages_expected !=ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after merging\n" ); goto unmap; }@@ -319,7 +250,7 @@ static void test_unmerge_zero_pages(void)/* Check if ksm_zero_pages is updated correctly after unmerging */ pages_expected /=2 ; -if (pages_expected !=get_my_ksm_zero_pages ()) { +if (pages_expected !=ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after unmerging\n" ); goto unmap; }@@ -329,7 +260,7 @@ static void test_unmerge_zero_pages(void) *((unsigned int *)&map[offs]) = offs;/* Now we should have no zeropages remaining. */ -if (get_my_ksm_zero_pages ()) { +if (ksm_get_self_zero_pages ()) {ksft_test_result_fail ("'ksm_zero_pages' updated after write fault\n" ); goto unmap; }tools/testing/selftests/mm/ksm_functional_tests.c static void init_global_file_handles(void) · line 693
Quoted 9 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-6@@ -685,19 +616,9 @@ static void init_global_file_handles(void) mem_fd =open ("/proc/self/mem" ,O_RDWR );if (mem_fd <0 )ksft_exit_fail_msg ("opening /proc/self/mem failed\n" ); - ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR ); -if (ksm_fd <0 ) -ksft_exit_skip ("open(\"/sys/kernel/mm/ksm/run\") failed\n" ); - ksm_full_scans_fd =open ("/sys/kernel/mm/ksm/full_scans" ,O_RDONLY ); -if (ksm_full_scans_fd <0 ) -ksft_exit_skip ("open(\"/sys/kernel/mm/ksm/full_scans\") failed\n" );For these skip cases, we should probably do a test access.
Quoted 7 lines
if (
ksm_stop_and_unmerge()< 0)ksft_exit_skip("accessing \"/sys/kernel/mm/ksm/run\\") failed\n"); if (ksm_get_full_scans() < 0) ksft_exit_skip("accessing \"/sys/kernel/mm/ksm/full_scans\") failed\n");So we later now that the files actually do exist and can be opened+written.
Agree
Quoted 5 lines
tools/testing/selftests/mm/ksm_functional_tests.c static void init_global_file_handles(void) · line 698
Quoted 5 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-2pagemap_fd = open("/proc/self/pagemap", O_RDONLY);> if (pagemap_fd < 0)ksft_exit_skip ("open(\"/proc/self/pagemap\") failed\n" ); - proc_self_ksm_stat_fd =open ("/proc/self/ksm_stat" ,O_RDONLY ); - proc_self_ksm_merging_pages_fd = open("/proc/self/ksm_merging_pages",
And I think we may do similar thing here.
if (ksm_get_self_merging_pages() < 0)
ksft_exit_skip("accessing \"/proc/self/ksm_merging_pages\") failed\n");Quoted 2 lines
tools/testing/selftests/mm/ksm_functional_tests.c static void init_global_file_handles(void) · line 700
Quoted 2 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-2- O_RDONLY); - ksm_use_zero_pages_fd =open ("/sys/kernel/mm/ksm/use_zero_pages" ,O_RDWR );
proc_self_ksm_stat_fd and ksm_use_zero_pages_fd is only used in test_unmerge_zero_pages(). So we can test and skip only there.
For proc_self_ksm_merging_pages_fd, it is used in all tests except test_prctl(). Check ahead maybe helpful.
Quoted 15 lines
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 559
Quoted 12 lines
tools/testing/selftests/mm/ksm_functional_tests.c +0
-0}int main (int argc,char **argv)tools/testing/selftests/mm/vm_util.c +2
-0@@ -555,3 +555,114 @@ bool detect_huge_zeropage(void)close (fd);return enabled; } + +long ksm_get_self_zero_pages ()For all these functions without parameters:
Quoted 2 lines
(void)
Thanks will add it.
Quoted 80 lines
Quoted 66 lines
tools/testing/selftests/mm/vm_util.c +66
-0+{ +int proc_self_ksm_stat_fd; +char buf[200 ]; +char *substr_ksm_zero; +size_t value_pos; +ssize_t read_size; +unsigned long my_ksm_zero_pages; + + proc_self_ksm_stat_fd =open ("/proc/self/ksm_stat" ,O_RDONLY ); +if (proc_self_ksm_stat_fd <0 ) +return 0 ; + + read_size =pread (proc_self_ksm_stat_fd, buf,sizeof (buf) -1 ,0 ); +close (proc_self_ksm_stat_fd); +if (read_size <0 ) +return -errno; + + buf[read_size] =0 ; + + substr_ksm_zero =strstr (buf,"ksm_zero_pages" ); +if (!substr_ksm_zero) +return 0 ; + + value_pos =strcspn (substr_ksm_zero,"0123456789" ); + my_ksm_zero_pages =strtol (substr_ksm_zero + value_pos,NULL ,10 ); + +return my_ksm_zero_pages; +} + +long ksm_get_self_merging_pages () +{ +int proc_self_ksm_merging_pages_fd; +char buf[10 ]; +ssize_t ret; + + proc_self_ksm_merging_pages_fd =open ("/proc/self/ksm_merging_pages" , +O_RDONLY ); +if (proc_self_ksm_merging_pages_fd <0 ) +return proc_self_ksm_merging_pages_fd; + + ret =pread (proc_self_ksm_merging_pages_fd, buf,sizeof (buf) -1 ,0 ); +close (proc_self_ksm_merging_pages_fd); +if (ret <=0 ) +return -errno; + buf[ret] =0 ; + +return strtol (buf,NULL ,10 ); +} + +long ksm_get_full_scans () +{ +int ksm_full_scans_fd; +char buf[10 ]; +ssize_t ret; + + ksm_full_scans_fd =open ("/sys/kernel/mm/ksm/full_scans" ,O_RDONLY ); +if (ksm_full_scans_fd <0 ) +return ksm_full_scans_fd; + + ret =pread (ksm_full_scans_fd, buf,sizeof (buf) -1 ,0 ); +close (ksm_full_scans_fd); +if (ret <=0 ) +return -errno; + buf[ret] =0 ; + +return strtol (buf,NULL ,10 );tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 635
Quoted 10 lines
tools/testing/selftests/mm/vm_util.c +10
-0+} + +int ksm_use_zero_pages () +{ +int ksm_use_zero_pages_fd; +ssize_t ret; + + ksm_use_zero_pages_fd =open ("/sys/kernel/mm/ksm/use_zero_pages" ,O_RDWR ); +if (ksm_use_zero_pages_fd <0 ) +return -1 ;For all these functions, we should not require the caller to lookup errno but instead return it.
Quoted 2 lines
return -errno;
Agree on the -errno thing, will change in all places.
Quoted 27 lines
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 639
Quoted 4 lines
tools/testing/selftests/mm/vm_util.c +4
-0+ + ret =write (ksm_use_zero_pages_fd,"1" ,1 ); +close (ksm_use_zero_pages_fd); +return ret ==1 ?0 : ret;return ret == 1 ? 0 : -errno;
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 649
Quoted 10 lines
tools/testing/selftests/mm/vm_util.c +10
-0+} + +int ksm_start_and_merge () +{ +int ksm_fd; +ssize_t ret; + + ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR ); +if (ksm_fd <0 ) +return -1 ;Same comments regarding errno.
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 653
Quoted 4 lines
tools/testing/selftests/mm/vm_util.c +4
-0+ + ret =write (ksm_fd,"1" ,1 ); +close (ksm_fd); +return ret ==1 ?0 : ret;double space
Quoted 17 lines
Quoted 4 lines
tools/testing/selftests/mm/vm_util.c +4
-0+} + +int ksm_stop_and_unmerge () +{tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 667
Quoted 10 lines
tools/testing/selftests/mm/vm_util.c +10
-0+int ksm_fd; +ssize_t ret; + + ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR ); +if (ksm_fd <0 ) +return -1 ; + + ret =write (ksm_fd,"2" ,1 ); +close (ksm_fd); +return ret ==1 ?0 : ret;double space
Shame on me :-(
Quoted 6 lines
Signature · 4 lines
-- Cheers, David / dhildenb
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH 2/3] selftests/mm: put general ksm operation into vm_util
On Thu, Jul 17, 2025 at 02:45:04AM +0000, Wei Yang wrote:
[...]
Quoted 10 lines
Looks reasonable, will remove it.
So how about leave three helpers:
ksm_merge()
ksm_start()
ksm_stop_and_unmerge()Would this be better?
Or maybe we just export ksm_merge() and ksm_unmerge(). Leave ksm_start() as an internal function, since only ksm_merge() use it.
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH 2/3] selftests/mm: put general ksm operation into vm_util
Quoted 7 lines · Wei Yang
On Thu, Jul 17, 2025 at 02:45:04AM +0000, Wei Yang wrote:
[...]Quoted 5 lines
Looks reasonable, will remove it.
So how about leave three helpers:
Quoted 8 lines
Quoted 5 lines
ksm_merge() ksm_start() ksm_stop_and_unmerge()Would this be better?
Or maybe we just export
ksm_merge()andksm_unmerge(). Leaveksm_start()as an internal function, since onlyksm_merge()use it.
Either that or start/stop. Whatever you think is best.
Signature · 3 lines
Cheers, David / dhildenb
Re: [PATCH 2/3] selftests/mm: put general ksm operation into vm_util
In reply to Wei YangQuoted 11 lines
My plan is there are two pairs of helper:
ksm_merge() <-> ksm_unmerge() ksm_start_and_merge() <-> ksm_stop_and_unmerge()Quoted 2 lines
Probably we should just use
ksm_stop_and_unmerge()and removeksm_unmerge().Looks reasonable, will remove it.
Quoted 7 lines
So how about leave three helpers:
ksm_merge() ksm_start() ksm_stop_and_unmerge()Would this be better?
Is there utility for separating ksm_start() and ksm_merge() in current code and with what you are intending to do?
Quoted 17 lines
Quoted 7 lines
Quoted 7 lines
tools/testing/selftests/mm/vm_util.c +7
-0+ +int ksm_stop_and_unmerge () +{ +int ksm_fd; +ssize_t ret; + + ksm_fd =open ("/sys/kernel/mm/ksm/run" ,O_RDWR );Quoted 9 lines
tools/testing/selftests/mm/vm_util.c bool detect_huge_zeropage(void) · line 667
Quoted 6 lines
tools/testing/selftests/mm/vm_util.c +6
-0+if (ksm_fd <0 ) +return -1 ; + + ret =write (ksm_fd,"2" ,1 ); +close (ksm_fd); +return ret ==1 ?0 : ret;double space
Quoted 1 line
Shame on me :-(
Not the end of the world ;)
Signature · 3 lines
Cheers, David / dhildenb
PATCH 3/3 selftests/mm: assert rmap behave as expected
In reply to Wei YangAs David suggested, currently we don't have a high level test case to verify the behavior of rmap. This patch introduce the verification on rmap by migration.
The general idea is if migrate one shared page between processes, this would be reflected in all related processes. Otherwise, we have problem in rmap.
Currently it covers following four scenarios:
* anonymous page
* shmem page
* pagecache page
* ksm page
---
v1: fault in region by just accessing it instead of print it
--- MAINTAINERS | 1 + tools/testing/selftests/mm/.gitignore | 1 + tools/testing/selftests/mm/Makefile | 3 + tools/testing/selftests/mm/rmap.c | 470 ++++++++++++++++++++++ tools/testing/selftests/mm/run_vmtests.sh | 4 + 5 files changed, 479 insertions(+) create mode 100644 tools/testing/selftests/mm/rmap.c
MAINTAINERS +1 -0
@@ -15917,6 +15917,7 @@ S: Maintained F: include/linux/rmap.h F: mm/page_vma_mapped.c F: mm/rmap.c +F: tools/testing/selftests/mm/rmap.c MEMORY MANAGEMENT - SECRETMEM M: Andrew Morton <[email protected]>
tools/testing/selftests/mm/.gitignore +1 -0
@@ -57,3 +57,4 @@ pkey_sighandler_tests_32 pkey_sighandler_tests_64 guard-regions merge +rmap
tools/testing/selftests/mm/Makefile +3 -0
@@ -100,6 +100,7 @@ TEST_GEN_FILES += hugetlb_dio TEST_GEN_FILES += droppable TEST_GEN_FILES += guard-regions TEST_GEN_FILES += merge +TEST_GEN_FILES += rmap ifneq ($(ARCH),arm64) TEST_GEN_FILES += soft-dirty
@@ -227,6 +228,8 @@ $(OUTPUT)/ksm_tests: LDLIBS += -lnuma $(OUTPUT)/migration: LDLIBS += -lnuma +$(OUTPUT)/rmap: LDLIBS += -lnuma + local_config.mk local_config.h: check_config.sh /bin/sh ./check_config.sh $(CC)
tools/testing/selftests/mm/rmap.c new file +470 -0
470 lines not on this page · raw message
tools/testing/selftests/mm/run_vmtests.sh +4 -0
@@ -83,6 +83,8 @@ separated by spaces: test handling of page fragment allocation and freeing - vma_merge test VMA merge cases behave as expected +- rmap + test rmap behave as expected example: ./run_vmtests.sh -t "hmm mmap ksm" EOF
@@ -527,6 +529,8 @@ CATEGORY="page_frag" run_test ./test_page_frag.sh aligned CATEGORY="page_frag" run_test ./test_page_frag.sh nonaligned +CATEGORY="rmap" run_test ./rmap + echo "SUMMARY: PASS=${count_pass} SKIP=${count_skip} FAIL=${count_fail}" | tap_prefix echo "1..${count_total}" | tap_output
Signature · 1 line
2.34.1
Quoted 231 lines · Wei Yang
As David suggested, currently we don't have a high level test case to verify the behavior of rmap. This patch introduce the verification on rmap by migration.
The general idea is if migrate one shared page between processes, this would be reflected in all related processes. Otherwise, we have problem in rmap.
Currently it covers following four scenarios:
* anonymous page
* shmem page
* pagecache page
* ksm pageSigned-off-by Wei Yang <[email protected]>Suggested-by David Hildenbrand <[email protected]>Cc David Hildenbrand <[email protected]>Cc Lorenzo Stoakes <[email protected]>Cc Rik van Riel <[email protected]>Cc Liam R. Howlett <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Harry Yoo <[email protected]>---
v1: fault in region by just accessing it instead of print it--- MAINTAINERS | 1 + tools/testing/selftests/mm/.gitignore | 1 + tools/testing/selftests/mm/Makefile | 3 + tools/testing/selftests/mm/rmap.c | 470 ++++++++++++++++++++++ tools/testing/selftests/mm/run_vmtests.sh | 4 + 5 files changed, 479 insertions(+)create mode 100644
tools/testing/selftests/mm/rmap.cMAINTAINERS +1
-0@@ -15917,6 +15917,7 @@ S: Maintained F: include/linux/rmap.h F: mm/page_vma_mapped.c F: mm/rmap.c +F: tools/testing/selftests/mm/rmap.c MEMORY MANAGEMENT - SECRETMEM M: Andrew Morton <[email protected]>tools/testing/selftests/mm/.gitignore +1
-0@@ -57,3 +57,4 @@ pkey_sighandler_tests_32 pkey_sighandler_tests_64 guard-regions merge +rmaptools/testing/selftests/mm/Makefile +3
-0@@ -100,6 +100,7 @@ TEST_GEN_FILES += hugetlb_dio TEST_GEN_FILES += droppable TEST_GEN_FILES += guard-regions TEST_GEN_FILES += merge +TEST_GEN_FILES += rmap ifneq ($(ARCH),arm64) TEST_GEN_FILES += soft-dirty@@ -227,6 +228,8 @@ $(OUTPUT)/ksm_tests: LDLIBS += -lnuma $(OUTPUT)/migration: LDLIBS += -lnuma +$(OUTPUT)/rmap: LDLIBS += -lnuma + local_config.mk local_config.h: check_config.sh /bin/sh ./check_config.sh $(CC)tools/testing/selftests/mm/rmap.c new file +148
-0@@ -0,0 +1,470 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* RMAP functional tests +* +* Author(s): Wei Yang <[email protected]> +*/ + +#include "../kselftest_harness.h" +#include <strings.h> +#include <pthread.h> +#include <numa.h> +#include <numaif.h> +#include <sys/mman.h> +#include <sys/prctl.h> +#include <sys/types.h> +#include <signal.h> +#include <time.h> +#include <sys/sem.h> +#include <unistd.h> +#include <fcntl.h> + +#include "vm_util.h" + +#define TOTAL_LEVEL 5 +#define MAX_CHILDREN 3 + +#define FAIL_ON_CHECK (1 << 0) +#define FAIL_ON_WORK (1 << 1) + +struct sembuf sem_wait = {0 ,-1 ,0 }; +struct sembuf sem_signal = {0 ,1 ,0 }; + +static const char initial_data[] ="Hello, world 0!" ; +static const char updated_data[] ="Hello, World 0!" ; + +enum backend_type { +ANON , +SHM , +NORM_FILE , +}; + +#define PREFIX "kst_rmap" +#define MAX_FILENAME_LEN 256 +const char *suffixes[] = { +"" , +"_shm" , +"_norm" , +}; + +struct global_data ; +typedef int (*work_fn )(struct global_data *data); +typedef int (*check_fn )(struct global_data *data); +typedef void (*prepare_fn )(struct global_data *data); + +struct global_data { +int worker_level; + +int semid; +int pipefd[2 ]; + +unsigned int mapsize; +unsigned int rand_seed; +char *region; + +prepare_fn do_prepare; +work_fn do_work; +check_fn do_check; + +enum backend_type backend; +char filename[MAX_FILENAME_LEN ]; + +unsigned long *expected_pfn; +}; + +/* +* Create a process tree with TOTAL_LEVEL height and at most MAX_CHILDREN +* children for each. +* +* It will randomly select one process as 'worker' process which will +* 'do_work' until all processes are created. And all other processes will +* wait until 'worker' finish its work. +*/ +int propagate_children (struct global_data *data) +{ +pid_t pid; +unsigned int num_child; +int status; +int ret =0 ; +int curr_child, worker_child; +int curr_level =1 ; +bool is_worker = true; + +repeat: + num_child =rand_r (&data->rand_seed) %MAX_CHILDREN +1 ; + worker_child = is_worker ?rand_r (&data->rand_seed) % num_child :-1 ; + +for (curr_child =0 ; curr_child < num_child; curr_child++) { + pid =fork (); + +if (pid <0 ) { +perror ("Error: fork\n" ); + }else if (pid ==0 ) { + curr_level++; + +if (curr_child != worker_child) + is_worker = false; + +if (curr_level ==TOTAL_LEVEL ) +break ; + + data->rand_seed += curr_child; + goto repeat; + } + } + +if (data->do_prepare) + data->do_prepare (data); + +close (data->pipefd[1 ]); + +if (is_worker && curr_level == data->worker_level) { +/* This is the worker process, first wait last process created */ +char buf; + +while (read (data->pipefd[0 ], &buf,1 ) >0 ) + ; + +if (data->do_work) + ret = data->do_work (data); + +/* Kick others */ +semctl (data->semid,0 ,IPC_RMID ); + }else { +/* Wait worker finish */ +semop (data->semid, &sem_wait,1 ); +if (data->do_check) + ret = data->do_check (data); + } + +/* Wait all child to quit */ +while (wait (&status) >0 ) { +if (WIFEXITED (status)) + ret |=WEXITSTATUS (status); + } + +return ret; +} +
tools/testing/selftests/mm/rmap.c line 158
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+FIXTURE (migrate) +{ +struct global_data data; +}; + +FIXTURE_SETUP (migrate) +{ +struct global_data *data = &self->data; + +ASSERT_EQ (numa_available (),0 );
if (numa_available() < 0)
SKIP(return, "NUMA not available");
Should that be a skip instead?
Quoted 56 lines
tools/testing/selftests/mm/rmap.c new file +56
-0+if (numa_bitmask_weight (numa_all_nodes_ptr) <=1 ) +SKIP (return,"Not enough NUMA nodes available" ); + + data->mapsize = getpagesize();> + +/* Prepare semaphore */ + data->semid =semget (IPC_PRIVATE ,1 ,0666 |IPC_CREAT ); +ASSERT_NE (data->semid,-1 ); +ASSERT_NE (semctl (data->semid,0 ,SETVAL ,0 ),-1 ); + +/* Prepare pipe */ +ASSERT_NE (pipe (data->pipefd),-1 ); + + data->rand_seed =time (NULL ); +srand (data->rand_seed); + + data->worker_level =rand () %TOTAL_LEVEL +1 ; + + data->do_prepare =NULL ; + data->do_work =NULL ; + data->do_check =NULL ; + + data->backend =ANON ; +}; + +FIXTURE_TEARDOWN (migrate) +{ +struct global_data *data = &self->data; + +if (data->region !=MAP_FAILED ) +munmap (data->region, data->mapsize); + data->region =MAP_FAILED ; +if (data->expected_pfn !=MAP_FAILED ) +munmap (data->expected_pfn,sizeof (unsigned long )); + data->expected_pfn =MAP_FAILED ; +semctl (data->semid,0 ,IPC_RMID ); + data->semid =-1 ; + +close (data->pipefd[0 ]); + + data->do_work =NULL ; + data->do_check =NULL ; + +switch (data->backend) { +case ANON : +break ; +case SHM : +shm_unlink (data->filename); +break ; +case NORM_FILE : +unlink (data->filename); +break ; + } +} + +int try_to_move_page (char *region) +{
tools/testing/selftests/mm/rmap.c line 225
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+int ret; +int node; +int status =0 ; +volatile unsigned long dummy =0 ; + +/* +* Fault in page in case it is not, otherwise move_pages() would +* return -ENOENT. +*/ + dummy = *((unsigned long *)region);
Use FORCE_READ() here
https://lkml.kernel.org/r/[email protected]
But, this really must happen in all children before actually performing the move in the worker. Otherwise the other processes don't map the page and will just ... fault it in later.
Quoted 11 lines
tools/testing/selftests/mm/rmap.c new file +11
-0+/* Prevent the compiler from optimizing out the entire loop: */ + asmvolatile ("" :"+r" (dummy)); + + ret =move_pages (0 ,1 , (void **)®ion,NULL , &status,MPOL_MF_MOVE_ALL ); +if (ret !=0 ) +return FAIL_ON_WORK ; + +/* Pick up a different target node */ +for (node =0 ; node <=numa_max_node (); node++) { +if (numa_bitmask_isbitset (numa_all_nodes_ptr, node) && node != status) +break ;
tools/testing/selftests/mm/rmap.c line 246
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ } + +if (node >numa_max_node ()) { +ksft_print_msg ("Couldn't find available numa node for testing\n" ); +return FAIL_ON_WORK ; + } + + ret =move_pages (0 ,1 , (void **)®ion, &node, &status,MPOL_MF_MOVE_ALL ); +if (ret !=0 ) +return FAIL_ON_WORK ;
Probably, if we don't manage to migrate, we should retry a couple of times and then SKIP.
Point is, migration might fail for various reasons (e.g., 2 NUMA nodes but one of them doesn't even have memory) etc.
Migration failures might indicate other problems, yes, but false failures from the test are suboptimal.
Quoted 17 lines
tools/testing/selftests/mm/rmap.c new file +17
-0+ +return 0 ; +} + +int move_and_update (struct global_data *data) +{ +int ret; + + ret =try_to_move_page (data->region); +if (ret !=0 ) +return ret; + +/* Change the content */ +strcpy (data->region, updated_data); + +return ret; +}
tools/testing/selftests/mm/rmap.c line 273
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ +int data_updated (struct global_data *data) +{ +if (data->region ==MAP_FAILED ) +return 0 ; + +if (strncmp ((char *)data->region, updated_data,strlen (updated_data))) +return FAIL_ON_CHECK ; +return 0 ; +}
I assume checking the PFN is sufficient. No need for the additional data content. In particular, with proper anon pages (CoW, see below) that doesn't work either way.
tools/testing/selftests/mm/rmap.c line 283
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ +TEST_F(migrate, anon) +{ +pid_t root_pid; +int ret; +struct global_data *data = &self->data; + +/* Map a shared area and fault in */ + data->region =mmap (0 , data->mapsize,PROT_READ |PROT_WRITE , +MAP_SHARED |MAP_ANONYMOUS ,-1 ,0 );
That is anon_shmem. We should test proper anon memory (MAP_PRIVATE), whereby pages are shared using CoW.
anon_shmem should behave mostly like shmem.
Quoted 19 lines
tools/testing/selftests/mm/rmap.c new file +19
-0+ASSERT_NE (data->region,MAP_FAILED ); +strcpy (data->region, initial_data); + + data->do_work = move_and_update; + data->do_check = data_updated; + + root_pid =getpid (); + + ret =propagate_children (data); + +if (getpid () == root_pid) { +if (ret &FAIL_ON_WORK ) +SKIP (return,"Failed on moving page" ); + +ASSERT_EQ (ret,0 ); + }else { +exit (ret); + } +}
[...]
tools/testing/selftests/mm/run_vmtests.sh separated by spaces: · line 87
Quoted 11 lines
tools/testing/selftests/mm/rmap.c new file +1
-0+TEST_HARNESS_MAIN tools/testing/selftests/mm/run_vmtests.sh +2
-0@@ -83,6 +83,8 @@ separated by spaces: test handling of page fragment allocation and freeing - vma_merge test VMA merge cases behave as expected +- rmap + test rmap behave as expected
"behaves"
Signature · 3 lines
Cheers, David / dhildenb
Quoted 1 line · David Hildenbrand
On 16.07.25 10:27, Wei Yang wrote:
[..]
Quoted 13 lines
tools/testing/selftests/mm/rmap.c line 158
Quoted 11 lines
tools/testing/selftests/mm/rmap.c new file +11
-0+ +FIXTURE (migrate) +{ +struct global_data data; +}; + +FIXTURE_SETUP (migrate) +{ +struct global_data *data = &self->data; + +ASSERT_EQ (numa_available (),0 );if (
numa_available()< 0)
Quoted 4 lines
SKIP(return, "NUMA not available");Should that be a skip instead?
You are right, will fix it.
Quoted 69 lines
Quoted 56 lines
tools/testing/selftests/mm/rmap.c new file +56
-0+if (numa_bitmask_weight (numa_all_nodes_ptr) <=1 ) +SKIP (return,"Not enough NUMA nodes available" ); + + data->mapsize = getpagesize();> + +/* Prepare semaphore */ + data->semid =semget (IPC_PRIVATE ,1 ,0666 |IPC_CREAT ); +ASSERT_NE (data->semid,-1 ); +ASSERT_NE (semctl (data->semid,0 ,SETVAL ,0 ),-1 ); + +/* Prepare pipe */ +ASSERT_NE (pipe (data->pipefd),-1 ); + + data->rand_seed =time (NULL ); +srand (data->rand_seed); + + data->worker_level =rand () %TOTAL_LEVEL +1 ; + + data->do_prepare =NULL ; + data->do_work =NULL ; + data->do_check =NULL ; + + data->backend =ANON ; +}; + +FIXTURE_TEARDOWN (migrate) +{ +struct global_data *data = &self->data; + +if (data->region !=MAP_FAILED ) +munmap (data->region, data->mapsize); + data->region =MAP_FAILED ; +if (data->expected_pfn !=MAP_FAILED ) +munmap (data->expected_pfn,sizeof (unsigned long )); + data->expected_pfn =MAP_FAILED ; +semctl (data->semid,0 ,IPC_RMID ); + data->semid =-1 ; + +close (data->pipefd[0 ]); + + data->do_work =NULL ; + data->do_check =NULL ; + +switch (data->backend) { +case ANON : +break ; +case SHM : +shm_unlink (data->filename); +break ; +case NORM_FILE : +unlink (data->filename); +break ; + } +} + +int try_to_move_page (char *region) +{tools/testing/selftests/mm/rmap.c line 225
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+int ret; +int node; +int status =0 ; +volatile unsigned long dummy =0 ; + +/* +* Fault in page in case it is not, otherwise move_pages() would +* return -ENOENT. +*/ + dummy = *((unsigned long *)region);Use
FORCE_READ()here
Quoted 6 lines
https://lkml.kernel.org/r/[email protected]
But, this really must happen in all children before actually performing the move in the worker. Otherwise the other processes don't map the page and will just ... fault it in later.
Ok, will access the region in all child before migrate.
Quoted 25 lines
Quoted 11 lines
tools/testing/selftests/mm/rmap.c new file +11
-0+/* Prevent the compiler from optimizing out the entire loop: */ + asmvolatile ("" :"+r" (dummy)); + + ret =move_pages (0 ,1 , (void **)®ion,NULL , &status,MPOL_MF_MOVE_ALL ); +if (ret !=0 ) +return FAIL_ON_WORK ; + +/* Pick up a different target node */ +for (node =0 ; node <=numa_max_node (); node++) { +if (numa_bitmask_isbitset (numa_all_nodes_ptr, node) && node != status) +break ;tools/testing/selftests/mm/rmap.c line 246
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ } + +if (node >numa_max_node ()) { +ksft_print_msg ("Couldn't find available numa node for testing\n" ); +return FAIL_ON_WORK ; + } + + ret =move_pages (0 ,1 , (void **)®ion, &node, &status,MPOL_MF_MOVE_ALL ); +if (ret !=0 ) +return FAIL_ON_WORK ;Probably, if we don't manage to migrate, we should retry a couple of times and then SKIP.
Quoted 6 lines
Point is, migration might fail for various reasons (e.g., 2 NUMA nodes but one of them doesn't even have memory) etc.
Migration failures might indicate other problems, yes, but false failures from the test are suboptimal.
Will add retry logic.
Quoted 32 lines
Quoted 17 lines
tools/testing/selftests/mm/rmap.c new file +17
-0+ +return 0 ; +} + +int move_and_update (struct global_data *data) +{ +int ret; + + ret =try_to_move_page (data->region); +if (ret !=0 ) +return ret; + +/* Change the content */ +strcpy (data->region, updated_data); + +return ret; +}tools/testing/selftests/mm/rmap.c line 273
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ +int data_updated (struct global_data *data) +{ +if (data->region ==MAP_FAILED ) +return 0 ; + +if (strncmp ((char *)data->region, updated_data,strlen (updated_data))) +return FAIL_ON_CHECK ; +return 0 ; +}I assume checking the PFN is sufficient. No need for the additional data content. In particular, with proper anon pages (CoW, see below) that doesn't work either way.
Quoted 14 lines
tools/testing/selftests/mm/rmap.c line 283
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ +TEST_F(migrate, anon) +{ +pid_t root_pid; +int ret; +struct global_data *data = &self->data; + +/* Map a shared area and fault in */ + data->region =mmap (0 , data->mapsize,PROT_READ |PROT_WRITE , +MAP_SHARED |MAP_ANONYMOUS ,-1 ,0 );That is
anon_shmem. We should test proper anon memory (MAP_PRIVATE), whereby pages are shared using CoW.
So the case should be
* mapping (MAP_PRIVATE | MAP_ANONYMOUS)
* write some content in root parent
* fault in for each child
* do migration and record pfn
* then check pfn is the same in each child
Quoted 42 lines
anon_shmemshould behave mostly like shmem.Quoted 19 lines
tools/testing/selftests/mm/rmap.c new file +19
-0+ASSERT_NE (data->region,MAP_FAILED ); +strcpy (data->region, initial_data); + + data->do_work = move_and_update; + data->do_check = data_updated; + + root_pid =getpid (); + + ret =propagate_children (data); + +if (getpid () == root_pid) { +if (ret &FAIL_ON_WORK ) +SKIP (return,"Failed on moving page" ); + +ASSERT_EQ (ret,0 ); + }else { +exit (ret); + } +}[...]
tools/testing/selftests/mm/run_vmtests.sh separated by spaces: · line 87
Quoted 11 lines
tools/testing/selftests/mm/rmap.c new file +1
-0+TEST_HARNESS_MAIN tools/testing/selftests/mm/run_vmtests.sh +2
-0@@ -83,6 +83,8 @@ separated by spaces: test handling of page fragment allocation and freeing - vma_merge test VMA merge cases behave as expected +- rmap + test rmap behave as expected"behaves"
Signature · 3 lines
Cheers, David / dhildenb
Signature · 2 lines
Wei Yang Help you, Help me
Quoted 182 lines · Wei Yang
Quoted 1 line · David Hildenbrand
On 16.07.25 10:27, Wei Yang wrote:
[..]
Quoted 13 lines
tools/testing/selftests/mm/rmap.c line 158
Quoted 11 lines
tools/testing/selftests/mm/rmap.c new file +11
-0+ +FIXTURE (migrate) +{ +struct global_data data; +}; + +FIXTURE_SETUP (migrate) +{ +struct global_data *data = &self->data; + +ASSERT_EQ (numa_available (),0 );if (
numa_available()< 0)Quoted 4 lines
SKIP(return, "NUMA not available");Should that be a skip instead?
You are right, will fix it.
Quoted 69 lines
Quoted 56 lines
tools/testing/selftests/mm/rmap.c new file +56
-0+if (numa_bitmask_weight (numa_all_nodes_ptr) <=1 ) +SKIP (return,"Not enough NUMA nodes available" ); + + data->mapsize = getpagesize();> + +/* Prepare semaphore */ + data->semid =semget (IPC_PRIVATE ,1 ,0666 |IPC_CREAT ); +ASSERT_NE (data->semid,-1 ); +ASSERT_NE (semctl (data->semid,0 ,SETVAL ,0 ),-1 ); + +/* Prepare pipe */ +ASSERT_NE (pipe (data->pipefd),-1 ); + + data->rand_seed =time (NULL ); +srand (data->rand_seed); + + data->worker_level =rand () %TOTAL_LEVEL +1 ; + + data->do_prepare =NULL ; + data->do_work =NULL ; + data->do_check =NULL ; + + data->backend =ANON ; +}; + +FIXTURE_TEARDOWN (migrate) +{ +struct global_data *data = &self->data; + +if (data->region !=MAP_FAILED ) +munmap (data->region, data->mapsize); + data->region =MAP_FAILED ; +if (data->expected_pfn !=MAP_FAILED ) +munmap (data->expected_pfn,sizeof (unsigned long )); + data->expected_pfn =MAP_FAILED ; +semctl (data->semid,0 ,IPC_RMID ); + data->semid =-1 ; + +close (data->pipefd[0 ]); + + data->do_work =NULL ; + data->do_check =NULL ; + +switch (data->backend) { +case ANON : +break ; +case SHM : +shm_unlink (data->filename); +break ; +case NORM_FILE : +unlink (data->filename); +break ; + } +} + +int try_to_move_page (char *region) +{tools/testing/selftests/mm/rmap.c line 225
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+int ret; +int node; +int status =0 ; +volatile unsigned long dummy =0 ; + +/* +* Fault in page in case it is not, otherwise move_pages() would +* return -ENOENT. +*/ + dummy = *((unsigned long *)region);Use
FORCE_READ()hereQuoted 6 lines
https://lkml.kernel.org/r/[email protected]
But, this really must happen in all children before actually performing the move in the worker. Otherwise the other processes don't map the page and will just ... fault it in later.
Ok, will access the region in all child before migrate.
Quoted 25 lines
Quoted 11 lines
tools/testing/selftests/mm/rmap.c new file +11
-0+/* Prevent the compiler from optimizing out the entire loop: */ + asmvolatile ("" :"+r" (dummy)); + + ret =move_pages (0 ,1 , (void **)®ion,NULL , &status,MPOL_MF_MOVE_ALL ); +if (ret !=0 ) +return FAIL_ON_WORK ; + +/* Pick up a different target node */ +for (node =0 ; node <=numa_max_node (); node++) { +if (numa_bitmask_isbitset (numa_all_nodes_ptr, node) && node != status) +break ;tools/testing/selftests/mm/rmap.c line 246
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ } + +if (node >numa_max_node ()) { +ksft_print_msg ("Couldn't find available numa node for testing\n" ); +return FAIL_ON_WORK ; + } + + ret =move_pages (0 ,1 , (void **)®ion, &node, &status,MPOL_MF_MOVE_ALL ); +if (ret !=0 ) +return FAIL_ON_WORK ;Probably, if we don't manage to migrate, we should retry a couple of times and then SKIP.
Quoted 6 lines
Point is, migration might fail for various reasons (e.g., 2 NUMA nodes but one of them doesn't even have memory) etc.
Migration failures might indicate other problems, yes, but false failures from the test are suboptimal.
Will add retry logic.
Quoted 32 lines
Quoted 17 lines
tools/testing/selftests/mm/rmap.c new file +17
-0+ +return 0 ; +} + +int move_and_update (struct global_data *data) +{ +int ret; + + ret =try_to_move_page (data->region); +if (ret !=0 ) +return ret; + +/* Change the content */ +strcpy (data->region, updated_data); + +return ret; +}tools/testing/selftests/mm/rmap.c line 273
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ +int data_updated (struct global_data *data) +{ +if (data->region ==MAP_FAILED ) +return 0 ; + +if (strncmp ((char *)data->region, updated_data,strlen (updated_data))) +return FAIL_ON_CHECK ; +return 0 ; +}I assume checking the PFN is sufficient. No need for the additional data content. In particular, with proper anon pages (CoW, see below) that doesn't work either way.
Quoted 14 lines
tools/testing/selftests/mm/rmap.c line 283
Quoted 10 lines
tools/testing/selftests/mm/rmap.c new file +10
-0+ +TEST_F(migrate, anon) +{ +pid_t root_pid; +int ret; +struct global_data *data = &self->data; + +/* Map a shared area and fault in */ + data->region =mmap (0 , data->mapsize,PROT_READ |PROT_WRITE , +MAP_SHARED |MAP_ANONYMOUS ,-1 ,0 );That is
anon_shmem. We should test proper anon memory (MAP_PRIVATE), whereby pages are shared using CoW.
Quoted 8 lines
So the case should be
* mapping (
MAP_PRIVATE|MAP_ANONYMOUS)
* write some content in root parent
* fault in for each child
* do migration and record pfn
* then check pfn is the same in each child
Hi, David
Is my understanding correct? If so, may I send a new version?
Signature · 2 lines
Wei Yang Help you, Help me
Sorry,
Quoted 5 lines
Quoted 5 lines
So the case should be
* mapping (
MAP_PRIVATE|MAP_ANONYMOUS)
* write some content in root parent
* fault in for each child
That is not required, but a FORCE_READ() would also not be harmful (if it would be done by your common code)
Quoted 2 lines
Quoted 2 lines
* do migration and record pfn
* then check pfn is the same in each child
Yes, makes sense.
A new version makes sense, so we can have a look what the new state looks like.
Signature · 3 lines
Cheers, David / dhildenb